Withdrawn: automated tests, described as passing

What was wrong before this entry: the site said there were over two thousand automated tests across the shipped systems, and it said they were passing.

The count survived audit. It is provable by a declaration census in each repository, and two thousand is well under every honest count anyone produced, which is why it is published as a floor rather than a figure.

The word passing did not survive. No repository in scope carries a committed run report — no JUnit XML, no test-results file, no coverage output. Nobody in the campaign had run the suites. A test that exists and a test that passes are different claims, and only one of them had an artifact behind it.

So the word came off and the count stayed. The path to earning it back is not an argument, it is a file: commit a dated run report per repository and the word is available again. Until then the honest sentence is the shorter one.

This entry is marked withdrawn rather than corrected because a claim that was public and is now not public should be findable. A quietly deleted claim teaches a reader nothing; a withdrawn one tells them exactly how this site treats a sentence it cannot support.